pengjunjie/ragflow
1
0
Fork 1
mirror of https://github.com/infiniflow/ragflow.git synced 2025-12-08 20:42:30 +08:00
Code Issues Packages Projects Releases Wiki Activity

refactor retieval_test, add SQl retrieval methods (#61)

Browse Source
This commit is contained in:
KevinHuSh
2024-02-08 17:01:01 +08:00
committed by GitHub
parent 0a903c7714
commit 5e0a689c43
16 changed files with 238 additions and 74 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
rag/nlp/search.py
View File

@ -268,9 +268,9 @@ class Dealer:
dim = len(sres.query_vector)
start_idx = (page - 1) * page_size
for i in idx:
ranks["total"] += 1
if sim[i] < similarity_threshold:
break
ranks["total"] += 1
start_idx -= 1
if start_idx >= 0:
continue
@ -280,6 +280,7 @@ class Dealer:
break
id = sres.ids[i]
dnm = sres.field[id]["docnm_kwd"]
did = sres.field[id]["doc_id"]
d = {
"chunk_id": id,
"content_ltks": sres.field[id]["content_ltks"],
@ -296,8 +297,9 @@ class Dealer:
}
ranks["chunks"].append(d)
if dnm not in ranks["doc_aggs"]:
ranks["doc_aggs"][dnm] = 0
ranks["doc_aggs"][dnm] += 1
ranks["doc_aggs"][dnm] = {"doc_id": did, "count": 0}
ranks["doc_aggs"][dnm]["count"] += 1
ranks["doc_aggs"] = [{"doc_name": k, "doc_id": v["doc_id"], "count": v["count"]} for k,v in sorted(ranks["doc_aggs"].items(), key=lambda x:x[1]["count"]*-1)]
return ranks